AI Is About to Crash. Here’s Why.

Hello world.I’m an unemployed ex Big Tech software engineer with 25 years of experience in the technology industry.And lately, I’ve been looking at the AI industry and thinking:Is it just me, or is this bubble looking extra bubbly?Because something feels different.The AI industry is still burning through enormous amounts of money. The companies at the center of the AI boom continue to require staggering amounts of capital to build data centers, buy GPUs, train models, and keep everything running.But suddenly, the tone seems to be changing.We’re seeing some of the most insanely valued companies in history preparing to go public. We’re hearing calls for government support. And we’re seeing AI executives who spent years predicting an imminent AI driven jobs apocalypse suddenly sounding a little more... cautious.Hmm.That’s interesting.To me, this looks a lot like peak bubble behavior.It reminds me of what happens near the end of every major financial mania. Everyone starts looking for the next investor willing to buy into the dream before the music stops.Or, as we like to call them in the financial world:Bag holders.But why is this happening now?I think there are several things happening beneath the surface that are starting to expose a fundamental problem with the current AI boom.And before the AI true believers come after me with pitchforks and flaming GPUs, let me be clear.I believe AI is going to be incredibly important.I just don’t believe we have Artificial General Intelligence.Not even close.We Don’t Have AGI. We Have Very Expensive Probabilistic Parrots.Yes, AI has gotten dramatically better.We have agentic systems. We have tool integrations. We have models that can write code, analyze documents, generate images, conduct research, and occasionally pretend to be a competent junior employee.But fundamentally, today’s large language models are still probabilistic systems predicting what comes next based on patterns learned from enormous amounts of data.They can do remarkable things.But they still struggle with consistency, reasoning, context, hallucinations, and knowing when they don’t know something.And because of that, I’m skeptical that today’s AI is suddenly going to invent the next Warp Drive and increase global productivity by 10,000 percent.Claude is not going to wake up tomorrow and announce:“Good morning, humans. I have solved faster than light travel. Also, I fixed the economy.”Sorry.But there is one thing AI could potentially do extremely well.It could replace human labor.And I think that’s the gigantic bet hiding underneath the entire AI boom.The AI Industry Has Made a Trillion Dollar Bet on Human LaborThink about the economics.The AI industry has attracted and consumed enormous amounts of capital.And whether we’re talking about investor money, corporate spending, or debt financing, the numbers are staggering.The basic premise behind much of this investment is that AI will eventually generate enormous profits by automating human work.Especially white collar work.And that’s the only economic story that really makes sense to me.Because if you are spending hundreds of billions, or potentially trillions, building the infrastructure required to run AI, eventually you need to make that money back.And you can’t do that by selling a $20 ChatGPT subscription to every person on Earth.You need something much bigger.You need to capture a meaningful percentage of the economic value currently generated by human workers.And that’s why, for years, we’ve heard predictions from some of the biggest names in AI about massive numbers of white collar jobs being automated.Software engineers.Lawyers.Accountants.Customer service representatives.Consultants.Researchers.Basically anyone who spends their day sitting in front of a computer.The logic is straightforward.If AI can replace millions of expensive human workers, the companies providing that AI can potentially capture a gigantic amount of economic value.And if you believe the more extreme AI predictions, we’re talking about tens of millions of workers eventually being replaced.That’s a massive opportunity.Except...There’s a problem.It’s not happening fast enough.The AI Productivity Revolution Is Taking Longer Than ExpectedFor all the talk about AI replacing human workers, we’re not seeing anything remotely close to the scale necessary to justify the industry’s enormous financial commitments.We’re certainly not seeing millions upon millions of white collar workers being replaced every